Individualized and Optimal Talent-Management of the AWF in Response to COVID-19: Dynamic Programming Approach

Abstract

This report is an extension of the originally proposed sequence of three studies that developed a cutting-edge modeling and simulation tool for the Acquisition Workforce (AWF). The initial objective of that sequence was to build a Dynamic Retention Model (DRM) from the ground-up for the AWF to restore and maintain a capable and flexible acquisition workforce in support of the needs of the modern warfighter. The current report uses the previous model to analyze the phenomenal and unprecedented impact of COVID-19 pandemic on the U.S. civilian sector and its potential effects on the size and composition of the AWF in the coming years. After going steadily down for almost a decade and being at the historical low of 3.5% in February 2020, the U.S. unemployment rate spiked to almost 15% in April 2020. This event represented an unparalleled increase of more than 11% in just two months. As surprising as the initial increase was the sharp fall in the U.S. unemployment rate that followed. As of November 2021, just a year and a half after the peak, the unemployment rate is hovering around 4.6%, barely more than one percentage point above the previous historical low. While the impact of COVID-19 has been so far much harsher on the civilian sector employment than on the government sector (i.e., and the AWF), it is unclear how the latter will evolve in the mid- and long-run after the fast, ongoing recovery of the private sector. We take advantage of the DRM developed in the previous studies and extend it to explore the potential consequences of economy-wide shocks (such as COVID-19) on the AWF as the economy shows signs of strong recovery. We start analyzing the behavior of a representative AWF worker at the beginning of the pandemic, when the strength of the economic recovery was highly uncertain. We find that, under a number of different scenarios regarding the speed of recovery, it takes several years (in expectation) before the AWF employee returns to the pre-pandemic behavior. The main effect of the COVID-19 shock is to make the AWF job temporarily more attractive than a similar job in the private sector, inducing the AWF worker to stay much longer in the government. A caveat of the previous analysis is that it assumes that the AWF employee is able to predict (in expectation) the recovery path of the economy. To address that unrealistic feature of the analysis, we extend the initial study by “forcing” the AWF worker to go through the strong economic recovery path observed after the outset of the pandemic. That is, we predict the agent behavior when the recovery paths are much more positive than originally forecasted. Not surprisingly, the initial higher valuation of the AWF job compared to the private sector quickly dissipates, and AWF attrition rates surge above pre-pandemic levels as employees who were planning to move to the private sector (and froze their plans due to the pandemic) resume their original courses of action. An important take-away is that, while the COVID-19 shock may initially induce more employees to stay longer in the AWF, it is not a permanent solution to retain valuable workers. To this end, traditional personnel policy actions will be required by the AWF leadership. We conclude the report by describing different possibilities to continue extending the model even further. These extensions will augment the DRM to provide the AWF leadership more accurate and powerful predictions of future AWF worker behavior.